ADP-ribosylation is a technically challenging PTM which has just emerged into the field of PTM-specific proteomics. But this fragile modifier requires special treatment on both a data acquisition and data processing level: it is highly labile under higher-energy collisional dissociation (HCD), and the degree of lability can depend on the site it modifies. Its behavior thus violates some assumptions on which proteomics algorithms are based. Here we present nonlocalized ADPr searching: a simple principle for maximizing sensitivity toward ADP-ribosylation when searching conventional HCD data. By scoring the strong fragment ions generally observed in ADPr spectra rather than the weak and often absent localization-dependent ions, nonlocalized searches are more sensitive. They also run significantly faster, due to reduced search space, and require no assumptions about which amino acids can be modified. We illustrate implementation in three search systems: Morpheus, MaxQuant, and MASCOT, and we also present a means of rapidly finding and extracting ADP-ribosylated peptide spectra from large datasets for more focused searching. This approach both improves identification of ADP-ribosylated peptides and avoids mis-localization of the modification sites.
Keywords: ADP-ribosylation; HCD fragmentation; Lability; Localization; Nonlocalized; Serine ADPr.